User talk:Aerik

Re: Wikidweb
In response to your comments on my talk page.


 * Wikipedia's spam policy is pretty straightforward on not having external links unless they're absolutely justified. If/when Wikidweb has a Wikipedia article, an internal link would be the most appropriate way to link to it from other Wikipedia articles.
 * My "spammy" comment is based on a perusal of wikidweb a couple of weeks ago when I first encountered it. http://wikidweb.com/wiki/Category:Men%27s_Health for instance, is full of spammy sites. Of course, "spam" has been co-opted by many online communities and means different things to different people. So, for you, it might be just what you're looking for.
 * As a dmoz editor, like most of my fellow editors, I try very hard to be objective when it comes to assessing web sites. We also don't mind a little competition and none of us who have been with the directory for any length of time would ever say that dmoz is perfect. But, it's the surfers we care about, not the web site owners. A lot of "perceived shortcomings" are a misunderstanding of dmoz's mission.

So, anyway, best of luck to you. --Wrathchild (talk) 15:54, 23 January 2006 (UTC)


 * I replied on his talk page --Aerik 23:09, 28 January 2006 (UTC)

Tsk!
You failed to mention that you are also an ODP editor. &mdash;Wrathchild (talk) 16:31, 13 February 2006 (UTC)

Trial category intersection
Hi there. Are you around and available to comment at Wikipedia talk:Category intersection? I've asked a few questions there about the trial system you've set up. It would be great if you could update to more recent data, more recent than November! Carcharoth 16:24, 21 January 2007 (UTC)


 * Responded to on the relevant pages :-) Sorry for not getting back to you sooner. --Aerik 18:21, 23 January 2007 (UTC)

Intersections
Just wondering if there is anything new to report about category intersections. I think there are many ways we could relieve the load on the servers if that becomes a problem. Also, I've written up a related proposal about a way to mine information from the wikilinks that already exist in articles. I hope you can take a look and give some feedback. -- Sam uel Wan t man 10:31, 7 April 2007 (UTC)

Copied from my (belated) email response:
 * Sorry to take so long to get back to you - I'd really hoped to have more prgress to report. I've been fooling around with Zend_Search_Lucene for testing, and I'm hitting some hurdles - not with the way I'm trying to do it, but due to the fact that I'm using a cheap host with limited RAM and cpu.


 * In short, I think Lucene is the way to go, and this can be implemented quite easily (comparatively). This is what Brion has been suggesting for a long time now, so I hope it will be pretty easy to get going once I've demonstated it... that's the hard part for now though... copying 1.5 million records into MySQL and indexing them is not exactly a quick and easy process... not that I'm complaining!  It's just taking me awhile.


 * Anyway, I think that implementing this will be quite simple, technically, given that we're already using Lucene for the main search index.

Also, regarding WP:LI - yes, that could be done in the same basic way as category intersections - cool idea, too. I'm working on setting up a fresh index on a local windows machine (plenty of RAM), and then I'll do some comparisons of MySQL's fulltext index vs. Lucene and report back. --Aerik 11:21, 15 April 2007 (UTC)

Zend Search Lucene
So, I'm messing with Zend Search Lucene - got it running on a couple of machines, my linux vps and an xp box at home. In each one it seems to freeze up after a certain amount of time. I don't get it. I either need to find a mailing list / forum somewhere, or maybe try pylucene. *sigh*... --Aerik 05:51, 18 April 2007 (UTC)